Writing about the NHS IT project feels a bit like beating a two-year-old at chess: it's easy to make yourself look clever, but only in an unreasonable context.

The best coverage of the national programme has been by Tony Collins in Computer Weekly; Michael Cross in The Guardian and the team led by Jon Hoeksma and Linda Davidson at E-Health Insider.

Looking at the most recent piece in The Guardian (linked to last night's broadcast of a More 4 News feature about the risks to NHS IT from computer viruses), I clicked through to Michael Cross's 2008 cogent analysis of the poverty of much of the debate about NPfIT and CFH.

Parts of the NHS IT programme - notably Picture Archiving and Communications Systems - have been great successes. The direction of travel in healthcare, with better use of data to show quality, outcomes and variations and address safety issues, is not fundamentally in question. It requires good IT.

Much of the problem stems from learning the wrong lessons from previous NHS IT fiascos in Wessex and the London Ambulance Service, which made the early years and design of NPfIT largely a procurement exercise, with too little thought for the end users and their wants and needs.

The late Bob Sang wrote about these issues for me in the March 2005 reform special issue of British Journal of Healthcare Management, outlining the NHS's need to move from centralised to distributed systems for IT. He did so from the perspective of having been involved in a major and successful telecoms project with Ericsson, as well as from his knowledge of management issues.

The need for a more localised NPfIT is a commonplace now. It wasn't then. That was Bob for you.